WildAi 2d294566b6 docs: refresh external loader docs for realtime ...

The fallback column named two families while four ship, and the
top changelog entry lagged the version. A test now keeps the
documented set in step with the loader table.

2026-10-06 01:50:32 +03:00

..

fixtures

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

__init__.py

refactor init

2026-08-14 20:26:31 +03:00

probe_aimdo_view_private_cost.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_core_slice_read_cost.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_dense_read_scale.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_external_fp8_full_path.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_file_to_vram_read_strategies.py

feat: load every weight route file->VRAM via core's aimdo DMA (v2.13.0)

2026-09-30 23:02:25 +03:00

probe_mapped_pages_uss.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_ram_census.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_safetensors_aliases_mapping.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_stream_reader_retention.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

probe_unified_dense_route.py

feat: load every weight route file->VRAM via core's aimdo DMA (v2.13.0)

2026-09-30 23:02:25 +03:00

probe_vbar_residency.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

realtime_e2e_support.py

feat: drop the dedicated realtime node, move to WMNodes categories (v2.10.0)

2026-09-27 01:45:18 +03:00

test_asr_external_native_load.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_asr_external_native.py

fix(load): register realtime packaged default

2026-10-06 01:49:22 +03:00

test_asr_generation.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_asr_loader.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_asr_node.py

feat: drop the dedicated realtime node, move to WMNodes categories (v2.10.0)

2026-09-27 01:45:18 +03:00

test_assign_loading.py

feat: auto-detect config_name, drop VibeVoice-Large (v2.7.0)

2026-08-27 20:33:19 +03:00

test_attention_utils.py

fix: shorten runtime warnings, drop deprecated gen-config pair

2026-10-04 01:48:47 +03:00

test_audio_backend.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_audio_utils.py

refactor(log): rename console prefix to [VibeVoice TTS]

2026-10-03 16:03:14 +03:00

test_base_loader.py

feat: load every weight route file->VRAM via core's aimdo DMA (v2.13.0)

2026-09-30 23:02:25 +03:00

test_bundle_release.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

test_cfg_loop.py

refactor init

2026-08-14 20:26:31 +03:00

test_comfy_stream.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_config_detect.py

fix(detect): add realtime embedding signature

2026-10-06 01:49:22 +03:00

test_config_name_normalization.py

fix(load): alias lowercase realtime spellings

2026-10-06 01:50:32 +03:00

test_conftest.py

refactor init

2026-08-14 20:26:31 +03:00

test_convrot_module.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_custom_types.py

feat: external model input via VibeVoiceLoadExternalModel node (v2.2.0)

2026-08-16 23:59:04 +03:00

test_decode_scaling_transform.py

sound fix

2026-08-14 20:49:06 +03:00

test_dense_1p5b_fixture_structure.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_device_utils.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_docs_consistency.py

docs: refresh external loader docs for realtime

2026-10-06 01:50:32 +03:00

test_double_load_holds_previous_model.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

test_dtype_attention_node_validation.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_dtype_utils.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_e2e_intelligibility.py

fix: cpu-first loading + neg-branch rope fix

2026-08-15 16:46:56 +03:00

test_e2e_lowvram_generation.py

feat: native lowvram streaming for oversized models (v2.6.0)

2026-08-27 17:52:17 +03:00

test_e2e_smoke_contract.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_ensure_downloaded_compat.py

fix: cpu-first loading + neg-branch rope fix

2026-08-15 16:46:56 +03:00

test_eos_calibration.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_eviction_on_change.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_extension.py

feat: drop the dedicated realtime node, move to WMNodes categories (v2.10.0)

2026-09-27 01:45:18 +03:00

test_external_loader_node.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_external_loader.py

fix(load): name the remedy in both errors

2026-10-06 01:50:32 +03:00

test_folder_registration.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_fp8_quant.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_generate_condition_shape.py

refactor init

2026-08-14 20:26:31 +03:00

test_generate_neg_position_ids.py

fix: version-safe config dtype access — silence torch_dtype deprecation (v2.7.1)

2026-08-27 21:49:17 +03:00

test_generate_progress_callback.py

fix: version-safe config dtype access — silence torch_dtype deprecation (v2.7.1)

2026-08-27 21:49:17 +03:00

test_generation_condition_source.py

fix: cpu-first loading + neg-branch rope fix

2026-08-15 16:46:56 +03:00

test_generation_family.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_generation.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_gguf_dequant_cache.py

perf: cache dequantized GGUF weights, closing ~35% of the q8_0 vs fp8 gap

2026-10-01 12:31:38 +03:00

test_gguf_dequant_perf.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_gguf_flat_pool.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_gguf_forward_diagnostics.py

refactor(log): rename console prefix to [VibeVoice TTS]

2026-10-03 16:03:14 +03:00

test_gguf_linear.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_gguf_loading.py

fix: list and load .gguf files in Load VibeVoice Model node (v2.2.1)

2026-08-17 01:29:12 +03:00

test_gguf_q8_0_fused.py

test: make the fused-kernel routing assertion non-vacuous

2026-10-01 12:47:50 +03:00

test_gguf_quant_blocks.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_gguf_quant_naming.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_imports.py

feat: external model input via VibeVoiceLoadExternalModel node (v2.2.0)

2026-08-16 23:59:04 +03:00

test_init.py

fix: lowvram partial-load completion, VibeVoice-7B rename follow-through, copy-free packaged tokenizer

2026-08-26 13:24:55 +03:00

test_integration.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_load_device_flow.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_loader_quant_paths.py

fix(loader): match the rooted encoder namespace

2026-10-06 01:49:23 +03:00

test_loader.py

feat: load every weight route file->VRAM via core's aimdo DMA (v2.13.0)

2026-09-30 23:02:25 +03:00

test_logging_idiom.py

refactor(log): rename console prefix to [VibeVoice TTS]

2026-10-03 16:03:14 +03:00

test_lowvram_protocol.py

feat: native lowvram streaming for oversized models (v2.6.0)

2026-08-27 17:52:17 +03:00

test_meta_init_feasibility.py

fix: version-safe config dtype access — silence torch_dtype deprecation (v2.7.1)

2026-08-27 21:49:17 +03:00

test_mm_api_surface.py

perf: stream sharded/quant loads, console progress, dep cleanup

2026-09-01 21:39:58 +03:00

test_model_forward.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_model_info.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_model_registry.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_no_v1_artifacts.py

refactor init

2026-08-14 20:26:31 +03:00

test_node_schema.py

feat: drop the dedicated realtime node, move to WMNodes categories (v2.10.0)

2026-09-27 01:45:18 +03:00

test_offload_contract.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_package_structure.py

refactor init

2026-08-14 20:26:31 +03:00

test_patcher_behavioral.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_patcher_weight_visibility.py

fix: cpu-first loading + neg-branch rope fix

2026-08-15 16:46:56 +03:00

test_patcher.py

feat: align VibeVoice loading with core DynamicVRAM, fix RAM ghost (v2.12.0)

2026-09-30 18:24:52 +03:00

test_pretrained_init_weights.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_processor_io_backend.py

feat: torchaudio-primary audio backend, librosa now optional (v2.1.0)

2026-08-15 21:54:53 +03:00

test_progress_utils.py

perf: stream sharded/quant loads, console progress, dep cleanup

2026-09-01 21:39:58 +03:00

test_pyproject.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_quant_e2e_parity.py

perf: fp8-resident weights + streaming safetensors load (v2.8.0)

2026-08-28 01:24:26 +03:00

test_quant_groundwork.py

feat: unload-on-change eviction + quant-resident runtime for GGUF and quantized safetensors (v2.4.0 -> v2.5.0)

2026-08-26 12:17:24 +03:00

test_ram_census.py

refactor(log): rename console prefix to [VibeVoice TTS]

2026-10-03 16:03:14 +03:00

test_ram_measurement_contract.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_ram_peak_sampling.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_ram_spike_regression.py

feat: unload-on-change eviction + quant-resident runtime for GGUF and quantized safetensors (v2.4.0 -> v2.5.0)

2026-08-26 12:17:24 +03:00

test_realtime_attention_parity.py

fix: shorten runtime warnings, drop deprecated gen-config pair

2026-10-04 01:48:47 +03:00

test_realtime_e2e_gpu.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_realtime_external_config.py

feat(config): ship packaged realtime config

2026-10-06 01:49:04 +03:00

test_realtime_generation_contract.py

fix: shorten runtime warnings, drop deprecated gen-config pair

2026-10-04 01:48:47 +03:00

test_realtime_generation.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_realtime_load_health.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_resident_streaming.py

feat: native lowvram streaming for oversized models (v2.6.0)

2026-08-27 17:52:17 +03:00

test_resolve_attention_kwarg.py

fix: cpu-first loading + neg-branch rope fix

2026-08-15 16:46:56 +03:00

test_sage_attention_patch.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_scheduler_index_regression.py

fix: version-safe config dtype access — silence torch_dtype deprecation (v2.7.1)

2026-08-27 21:49:17 +03:00

test_streaming_cache_compat.py

test: clean the suite to match the current architecture, and de-localise it

2026-10-01 15:49:15 +03:00

test_streaming_progress_callback.py

feat: report standard ComfyUI progress during TTS/ASR inference (v2.1.1)

2026-08-16 02:24:52 +03:00

test_streaming_progress.py

fix: correct realtime VibeVoice output on transformers 5.3

2026-09-26 23:30:37 +03:00

test_streaming_quant_load.py

test(quant): cover the realtime quant scan

2026-10-06 01:50:32 +03:00

test_tie_weights_compat.py

refactor init

2026-08-14 20:26:31 +03:00

test_timestep_embedder.py

perf: fp8-resident weights + streaming safetensors load (v2.8.0)

2026-08-28 01:24:26 +03:00

test_unified_tts_node.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_version_sync.py

feat: streaming quant load, single-rounding dequant, release hygiene (v2.11.0)

2026-09-28 18:25:19 +03:00

test_voice_presets.py

refactor: log through ComfyUI's own root-logger idiom, re-audit every level

2026-10-01 23:08:04 +03:00

test_workflow.py

fix: lowvram partial-load completion, VibeVoice-7B rename follow-through, copy-free packaged tokenizer

2026-08-26 13:24:55 +03:00